How do you make a client and server in docker

I’m trying out tensorflow serving in docker and i’m not sure how you create a client and server. Do I have to create a duplicate of the container with serving ?
Is there a better way to do this.
Some code samples would be helpful